Why most academy video setups fail
Most academies have a video story that goes like this: a senior coach films a few clips on their phone during a session, the videos sit in their camera roll, occasionally a clip gets sent to a parent on WhatsApp, and the rest is lost. There is no library, no searchability, no analysis, and no continuity. When the senior coach leaves, the video record of the academy goes with them.
Specialist video tools exist (Hudl, Dartfish, the various pro-grade systems) but they sit outside the academy management system. Coaches have to re-key player names, manual tag everything, and remember to upload — which means in practice almost nothing makes it in. The integration cost is the killer.
Cricket-specific AI, in the same login
FlikVault includes video AI as a native module. A coach films a clip on their phone in the FlikVault app, picks the player it's for, and uploads. The cloud processing runs ball tracking, pose estimation, slow-motion segmentation, and shot/delivery classification. The processed clip lives against the player in the academy database — searchable, taggable, attached to the session, and ready to embed in the next report.
Because the same identity system handles the player record across attendance, billing, and reports, every clip carries the right metadata automatically. There's no re-keying, no “which kid was this again,” no orphan files.
What the AI actually does
FlikVault's video pipeline is built specifically for cricket. It includes:
- Ball tracking. Detects and traces the ball through the delivery and shot, including pitch map and flight metrics where the camera angle supports it.
- Pose estimation. Frame-by-frame skeletal tracking of the player to analyse stance, trigger movement, head position, balance, follow-through.
- Slow-motion segmentation. Automatic identification of the key moment (delivery release, point of impact, foot landing) with smooth slow-mo around it.
- Shot and delivery classification. The system tags clips by shot type for batters and delivery type for bowlers, building a searchable library.
- AI coach notes. An AI agent — using cricket-specific prompts and access to the player's history — drafts coaching observations the coach can accept, edit, or discard. The coach is always in the loop; the AI is a starting point, not the authority.
Powered by SKrutin, sister product to FlikVault
The video AI capability comes from SKrutin, our dedicated cricket vision engine. SKrutin is the result of years of development on cricket-specific computer vision and AI-coaching models. FlikVault calls it as a service, which means academy customers get the benefit of ongoing model improvements without any integration work — the upgrades just appear.
Other academies and federations license SKrutin separately. Inside FlikVault, it's included.
Sharing, with the right permissions
Coaches can share clips with the player and parents through the portal. Clips can be embedded in term reports. Multi-clip comparisons (the player's shot in week one vs. week ten) can be assembled in a few taps. Sharing is permission-aware — directors configure whether parents see all clips, only the ones flagged for them, or only those embedded in reports.
Storage and bandwidth, handled
Cricket academies generate a lot of video. FlikVault uses cloud object storage with CDN delivery, so playback is smooth from anywhere, on phones or laptops, regardless of academy site location. Your IT footprint is zero — no NAS, no on-site server, no coach's phone running out of space.
Compliance and consent, by default
Video of minors is sensitive. FlikVault treats consent as a first-class object: parents sign off on video capture and on sharing scope as part of intake, and that consent is preserved against every clip. If a parent revokes consent, the system enforces it. The audit trail tells you which clips were seen by whom and when.
This is what ‘AI for cricket’ should mean
The phrase “AI for cricket” gets used loosely. In FlikVault it means something specific: cricket-trained models running on cricket-shaped data, attached to the right player record, available where you actually need it. Not a chatbot. Not a bolted-on plugin. Not a separate app. The vision engine and the management system understand each other natively, because they're built by the same team.